Need a book? Engineering books recommendations...

Return to index: [Subject] [Thread] [Date] [Author]

Re: Enercalc's Structural Engineering Library Survey

[Subject Prev][Subject Next][Thread Prev][Thread Next]
Bill raises a very interesting question. It is not simply whether we would
want to use a spreadsheet program or a compiled program. Personally, I have
been writting spreadsheet programs since SuperCalc in CPM format. I believe
the real issue is whether or not the program that we use yields valid
results. The level of conservatism in the results is a very subjective
opinion - for which I would lean to the conservative side for fear for
having to trade my home in order to defend my opinon against an expert
witness pissing contest (excuse the venacular).
I try to choose programs that fill my needs. Again, the answer to what I
would use has much to do with trust in the output.  We all know that
"virtually" no program is free from bugs. Imagine having the ability to take
a well known program and then modify it for your personal style and
inadvertently introducing yet another mistake into your new spreadsheet.
None of us are infallable - as many of you pointed out with my recently
released lateral analysis spreadsheet on the SEAOC web.
What control does a plan checker have when an analysis from a modified
version of Enercalc should appear on his desk. Assuming that an error has
been added by the user, who is responsible to find this error? Now many of
you may flame me by saying that plan checkers rely upon the results in
commercial programs - unquestioned. But I believe that they do (to a
degree), if only because they rely upon the reputation and reliability of
the program from many submittals or from owning and using the program
themselves. Nothing, however, can beat professional intuition based upon
experience and understanding how the materials you are designing work in the
real world.
Years ago I wrote a spreadsheet based program called Equake - for the design
of URM structures. Los Angeles City Plan check was the best beta team simply
because they nit-picked the program to death to understand not only the
results but to confirm that I, as the author of the program, understood the
design methodology and have translated it correctly to spreadsheet. I
believe too that the debates where a great learning experience for us all as
we debated the methodology to instill it better in each of our minds. Then
there was the debates as to how some of the analysis needed to be averaged
in order to make it work on multiple story structures with irregularities in
their diaphragms. Finally after almost a year and about one hundred
submittals using the program did the building department start to have some
faith in the programs results. I believe that when companies such as
Enercalc first introduced their programs to the market, they had to provide
more hand analysis problems to compare against their spreadsheet results - I
know I did with Equake. I included sample comparisons in my manual - which
brings up yet another topic.
Manuals explain a programs design assumptions. They help the user to
understand how their input is evaluated.  When I wrote Equake I was assulted
(verbally) by a young engineer in my office that assumed the input requested
(in this case, tributary area's) were figured in the same way that she
regularly performed the analysis manually. This was not the case since much
of the ability of the program was based upon letting the software compile
information which the user input level for level. This particular engineer
input vaules based upon adding one level to the level above. When she found
out that she had to redo a few building because we either had not written a
manual or she made an assumption without asking the programmer - well, she
literally hit the roof and unfortunately I was at the receiving end.
I have not done much contract plan checking in the last year, but when I did
I saw many proprietary softwares developed in spreadsheet, basic, visual
basic, and just about every other programing language out there. In almost
all of the cases, the input was listed in one area and the output in
another. The path between the two was very obscure or non-existent. Only
programs like MathCad yield results comparable to hand analysis. I can't
tell you how many hours were wasted when a calc did not seem to be yielding
the correct results (from experience gained intuition) and having to
manually go through a calc to find out that the users program was at fault.
Finally, I guess that I would have to place my vote for the compiled version
of Enercalc and trusting my ability to find an error and report it to the
manufacturer for correction. I would hope that the software vendor would
respond as quickly as possible to correct the mistake. I am still waiting
for many software companies to correct errors - Enercalc is not one of them
unless the error is something benign. I have found Enercalc willing to
embrace constructive criticism and use it to thier advantage. I have heard
many of the complaints, and even though I have known Mike and Hugh Brooks
for many years, my opinons are one of high regard for their ability to
create useful tools for engineers, be concerned about the validity of their
software and to be concerned about what their users think about their
products. Enercalc made my life easier and produced tools that yield results
(for the most part) that are in line with my methods of analysis. No one is
infallable, no program is bug free - the test of a good program is their
tech support response and their ability to correct a mistake in a timely
fashion.
In case I got off the track to defend Enercalc, my point is that we can not
contol the validity in the output of commercial programs that are allowed to
be modified by the end user. We can, however, hope that the software
developer who wishes to continue to serve a specified market, is willing to
listen to the end user and provide the tools that the majority of uses want.
If a users needs are more specific or do not represent the method of
analysis performed by the weighted average of professionals, he needs to
think about creating his own program from scratch and be prepared to prove
the validity of his results.

Dennis Wish PE